EMDR Therapy Quarterly

Introduction

EMDR is a transdiagnostic, integrative psychotherapeutic approach for dealing with memories of adverse life experiences (Farrell et al., 2020). As a treatment intervention, it is empirically validated and recommended as a treatment of choice for post traumatic stress disorder (PTSD) by leading organisations such as the World Health Organisation (2013); US Department of Veterans Affairs/ Department of Defence (2017); International Society for Traumatic Stress Studies (2018); and the National Institute for Health and Care Excellence (2018).  

EMDR therapy is designed to alleviate the distress caused by intrusive and unwanted memories of traumatic events (Landin-Romero et al., 2018). Using a structured 8-phase protocol and dual attention tasks, usually bilateral, saccadic eye movements (EMs), clients are guided to activate the distressing memory as an image, along with the body sensations, emotions, and negative beliefs they hold about themselves and follow the therapist’s moving hand while the brain processes the information. New associations can then be made with adaptive information already held in other memory networks (Shapiro, 2014). Shapiro proposed that EMs facilitate information processing when utilised in EMDR therapy and originally described them as a “crucial element” of EMDR (Shapiro, 1989, p.220). She revised this view when evidence emerged that other dual attention tasks, such as auditory bleeps and tapping, were also effective (Andrade, Kavanagh, & Baddeley, 1997; van den Hout et al., 2001; Servan-Schreiber et al., 2006).  

The core theoretical model of EMDR is that of adaptive information processing (AIP) which proposes an inherent information processing system that integrates new experiences into existing memory networks. These networks already hold the knowledge and information necessary to make sense of our experiences. Adverse life experiences and distressing incidents cause an imbalance in the nervous system, resulting in incomplete processing. The experience is then stored in a state-specific form as trauma memories. These distressing and dysfunctional memories are easily triggered by internal and external cues and can result in negative emotional, cognitive, and behavioural reactions (Soloman, & Shapiro, 2008).  

The hypothesised working mechanism of EMDR is still being investigated, but most evidence is gained for the working memory account (De Jongh et al., 2020).  As Matthijssen et al. (2021) indicate, working memory has limited capacity and that performing one task limits the performance of another. When this is applied to trauma memories, using bilateral stimulation (BLS) these dual tasks compete with working memory capacity with a consequence that the emotional intensity of the memory is gradually lost and eventually reconsolidated in a less emotional and distressing form. 

Despite international endorsement for EMDR therapy, its mechanism of action, and in particular the specific role of EMs as an inherent part of the intervention, continues to promote debate and critical consideration (Houben et al.,2018,2020; Kenchel et al., 2020; Leer et al., 2014, 2020; Ost & Easton, 2006). As Landin-Romero et al. (2018) suggests, since its inception, EMDR has instigated controversy surrounding the active ingredient of the treatment and whether the underlying mechanism differs from standard exposure or other trauma-focused treatment interventions. 

Landin-Romero et al. (2018) conducted a comprehensive, systematic review of the mechanism of action of EMDR therapy which included 87 studies. The authors determined that EMs provided a faster effect and had a moderate but significant additional treatment effect gain than other forms of bilateral stimulation. The Baek et al. (2019) study found that bilateral stimulation when compared with control studies, led to a clear and persistent decrease in fear behaviour. Additionally, they concluded that the BLS increased neuronal activity in the superior colliculus and the mediodorsal thalamus, thus dampening the excitability of neurons in the basolateral nucleus of the amygdala. 

The purpose of this paper was to provide a narrative systematic review of the literature since Landin-Romero et al. (2018). 

Methods

A review of the literature was examined using PubMed, PsychINFO, Web of Knowledge, Science Direct, Cochrane and Scopus databases. The search dates were between January 2018 and May 2021. Search terms included “eye movements or EMs”, “bilateral stimulation or BLS”, “working memory”, and “eye movement desensitisation and reprocessing or EMDR” with the Boolean connectors “or” and “and” applied.  The final selection of articles included used the following criteria:

• Original articles published in peer-reviewed journals, 
• Adult populations, 
• English language,  
• Conducted in either a therapy context or laboratory studies,  
• Published since January 2018.  

Figure 1 shows the PRISMA 2020 flow diagram for new systematic reviews (Page et al., 2020). .  

A total of six studies met the inclusion criteria (see Table 1).

Author & year	Sample (n) & population	Main findings	Conclusion
Eğeci & Özgün, 2019	100  Non-clinical	R-TEP can be successfully applied immediately following an inescapable event, and that EMs are an important and necessary component of EMDR treatment.	EMs reduce the vividness and emotionality of aversive memories
Van Schie, van Veen & Hagenaars, 2019	250  Non-clinical   University students	Despite WM taxation, the DA tasks did not differ from the control conditions for vividness or emotionality. When the intervention was longer, the result was more reduction in the number of intrusions than the control.	EMs do not affect the vividness or emotionality of aversive memories.
Littel & van Schie. 2019	44  Non-clinical  University students	There is a linear relationship between the amount of WM taxation and memory degradation. There is evidence that there may be an optimal level of WM taxation.	EMs reduce the vividness and emotionality of aversive memories.
Mertens et al., 2019	100  Non-clinical  University students	EMs reduce the vividness of negative autobiographical memories but not the emotionality.	EMs reduce the vividness of aversive memories but not the emotionality.
Bartels et al., 2018	80  Non-clinical  University students	EMs reduce the self-reported emotionality, vividness and arousability of experience-based and imagination-based sexual fantasies.	EMs reduce the vividness, emotionality and arousability of sexual images.
Homer & Deeprose, 2018	38  Non-clinical  University students	EMs reduce the vividness of voluntary images in social anxiety, but intrusive and involuntary images appear more resilient to EM attenuation.	EMs reduce the vividness of aversive memories but not the emotionality.

Results

The Homer and Deeprose (2018) laboratory study utilised a small pilot population to explore the EM paradigm beyond PTSD to individuals with social anxiety.  The study was designed to test whether EMs reduced the vividness and/or emotionality of intrusive social-anxiety images in a sub-clinical population. Thirty-eight non-clinical university students with a mean age of 20 years and self-reported intrusive social-anxiety images were recruited and assigned to either an EMs or control condition.  The EMs group visualised their intrusive social-anxiety image whilst concurrently responding to a visual task where letters were flashed on alternate sides of a computer screen on a striped background. The control group did the same, but the letters were in the centre of the screen on a white background and required no EMs. The results of the study showed that EMs reduced the vividness of autobiographical memories but not the emotionality.   

The Bartels et al. (2018) study was ostensibly a proof-of-concept investigation aimed to test whether EMs impair the vividness and/or emotionality of both experience-based and imagination-based sexual fantasies in a non-clinical sample. Eighty student participants were recruited and participated in all four conditions: Experience-based sexual fantasy image + EMs, Imagination-based sexual fantasy image + EMs, Experience-based sexual fantasy image + No EMs and Imagination-based sexual fantasy image + No EMs. The EMs were made by following a white dot on a computer screen for 96 EMs. After the EMs task, participants visualised the same sexual fantasy and re-rated its emotionality, vividness, and arousability. This procedure was completed for both an experience-based fantasy and an imagination-based fantasy.  The study results showed that self-reported vividness, emotionality and arousability decreased for the EMs condition across all participants regardless of whether the fantasies were experience-based or imagination-based. Vividness ratings increased for the No-EMs group.  

The third study, by Mertens et al. 2019, compared EM tasks commonly used in lab studies to see which reduced the emotionality and vividness of negative emotional memories.  They aimed to directly compare the three EMs tasks often used in testing working memory theory in laboratory research. One hundred non-clinical participants with self-reported negative autobiographical memories were randomly allocated to two conditions. In the first, they recalled one negative autobiographical memory whilst performing a control task. In the second, they recalled another negative autobiographical memory whilst performing one of three EMs tasks, dot tracking (Gunter & Bodner, 2008), letter identification (Andrade et al., 1997), or dot tracking + letter identification. Mertens et al. 2019, observed a slight reduction of memory vividness but no significant reduction in emotionality, and they also did not observe any correlation between the amount/type of working memory taxation and reduction of memory emotionality/vividness. Interestingly, they also found that some participants’ emotionality and vividness increased. These findings are inconsistent with many other studies showing a more substantial effect of tasks that are more taxing of the working memory and question whether an inverted U-curve, with neither too little nor too much taxation on the working memory, could explain why there is an increase in emotionality. Littel and van Schie explore this in their 2019 replication study. 

Littel and van Schie (2019) carried out an extended replication of Engelhard et al. (2011) study, where participants recalled four aversive autobiographical memories that were subjected to four levels of dual-attention working-memory taxation. They extended the study by adding a 24 hour follow-up to assess the longer-term change of degraded memories. The study aimed to research whether there is a dose-response relationship between working-memory taxation and memory degradation. They wanted to know if it is linear, whereby the more taxing a dual attention task is, the more significant the reduction in vividness/emotionality, or if it is an inverted U-curve. An inverted U-curve in this context suggests that too little taxation results in not enough competition for working memory resources or, alternatively, where too much taxation results in too much competition for resources, which leaves insufficient room for memory recall. Forty-four university students who experienced self-reported aversive autobiographical memories were instructed to recall a memory, perform the required task, and rate the memory for vividness and emotionality, and in addition, utilised a 24 hour follow-up. 

The researchers subjected the results to Bayesian analysis and assessed the likelihood of a linear or inverted U-curve relationship. The results showed a linear dose relationship between working-memory taxation and memory degradation immediately after the intervention and some unconvincing evidence for this pattern after one day. They found no evidence for a U-curve suggesting that there is potentially an optimal level of working-memory taxation. 

Eğeci and Özgün (2019) conducted a randomised control analogue study with laboratory-induced learned helplessness (LH). The study aimed to establish the efficacy of EMDR early intervention protocols using the Recent Traumatic Experiences Protocol (R-TEP) to see if EMDR can reverse the effects of induced learned helplessness (LH). One hundred participants with laboratory-induced LH were randomly assigned to one of five groups of equal numbers (20). The control group were given solvable maze puzzles, whilst the remaining four groups were given unsolvable maze puzzles and then received either full R-TEP, R-TEP with no EMs, LH + Wait 15 or LH + No Wait. The results found that R-TEP with EMs was significantly more effective than R-TEP with no EMs or control, which was essentially exposure only. As a result of these findings, the authors concluded that their study demonstrates that EMs reduce memory vividness and emotionality and may facilitate cognitive flexibility.  

The van Schie, van Veen, and Hagenaars (2019) study involved an extended replication examining whether dual attention tasks reduced the vividness/emotionality of newly formed trauma memories which they extended to include a novel hypothesis of whether those decreases in vividness/emotionality predicted lower frequency of intrusive memories. They used voluntary and involuntary memories to replicate clinical situations. In the largest study reviewed 250 university students were shown a nine-minute excerpt of a film that depicted a brutal murder with extreme violence that had been used in earlier studies (Nixon et al., 2009; Verwoerd, de Jong, & Wessel, 2008; Verwoerd et al., 2009, 2011; et al., Weidmann et al., 2009). In the first experiment, 76 participants were allocated to three groups, Recall + EMs, Recall + Counting or No-Task control. Surprisingly, the results showed that dual attention tasks did not decrease the vividness or emotionality for participants. This unexpected result led to another experiment with prolonged intervention times. In experiment two, 74 participants were allocated to one of the same three conditions as experiment one with a task intervention of 10 minutes. Although the results showed that participants experienced fewer intrusions, the dual attention task did not affect memory vividness or emotionality. To address the limitations identified in experiment two, a third and final experiment was undertaken. This experiment incorporated a recall-only control, and 100 participants were equally divided into this or the other three groups. The results of experiment three did not replicate those in experiment two, and the authors reported inconclusive results and questioned whether other mechanisms were at play.   

A broader perspective of the results from this systematic review highlights that EMs have been found to reduce the vividness of aversive memories in five studies (Homer & Deeprose, 2018; Bartels et al., 2018; Mertens et al., 2019; Littel & van Schie, 2019 and Eğeci & Özgün, 2019) as well as the emotionality of aversive memories in three of the same studies (Bartels et al., 2018; Littel & van Schie, 2019 and Eğeci & Özgün, 2019). Interestingly, Mertens et al. (2019) espoused that vividness and emotionality of some aversive memories increased for some participants and suggest that this is a result of the “changes in memory accessibility due to repeated recollection” (Mertens et al., 2019, p. 303). A contrary perspective is provided by van Schie et al. (2019), who concluded that EMs did not affect either vividness or emotionality.   

Discussion

This systematic review aimed to explore the current literature regarding the contribution of EMs to EMDR therapy. Two previous systematic reviews (Jeffries & Davies, 2013; Landin-Romero et al., 2018) and one meta-analysis (Lee & Cuijpers, 2013) summarised the literature until 2018. In their review, Jeffries and Davies (2013) recommended that large-scale dismantling studies were needed to reliably examine the difference in outcome between EMDR with EMs and EMDR without. The search conducted for this review, for clinical trials comparing EMDR with EMs to EMDR without EMs, completed since 2013, found that there had been two trials completed (Sack et al., 2016; Schubert, Lee & Drummond, 2016).  The remaining studies that have looked at the contribution of EMs in EMDR therapy have been laboratory studies with non-clinical populations and are the subject of this review (Homer & Deeprose, 2018; Bartels et al., 2018; Mertens et al., 2019; Littel & van Schie, 2019; van Schie et al., 2019; Eğeci & Özgün, 2019). These studies have concluded that EMs reduce the vividness and emotionality of aversive memories in healthy university student participants. The exception to this comes from the study by van Schie et al. (2019), which reported inconclusive results.  

The most robust of the studies reviewed, in terms of the hierarchy of evidence, is by Eğeci and Özgün (2019). This study consists of a large sample size (n =100) and contains a rigorous research methodology and structure with strong integrity in RCT (randomised controlled trials) design and subsequent application. The disadvantage, however, as with most of the studies in this systematic review, is that research participants are not clinical populations which, by corollary, questions the replication of results within the clinical environment. Secondly, there is no adherence to the specific attributes of EMDR therapy as an eight-phase intervention, which again raises concerns regarding clinical applicability. 

The study by van Schie et al.  (2019) had the largest sample (n = 250) and was an extended replication study. It utilised a frequently used paradigm for investigating intrusive memory development. However, it neither utilised an EMDR protocol nor did the participants have negative autobiographical memories. The intrusive memories were laboratory-induced, although the study did use voluntary and involuntary memories to replicate clinical PTSD. The researchers aimed to test whether dual attention tasks reduced the vividness and/or emotionality of induced trauma memories and used Ems. However, their results were inconclusive, and therefore the study is limited in its ability to generalise to clinical practice.   

Mertens et al. (2019) also utilised a large sample population (n = 100) and tested three EM tasks. The design incorporated a within-subject and between-subject procedure, and participants were randomly allocated to the conditions. This study used participants’ negative autobiographical memories, and despite a substantial working memory taxation, they observed a small reduction in memory vividness and no significant reduction in emotionality. They acknowledge that these findings are inconsistent with both working-memory theory and the results of many prior studies and suggest a possible inverted U-curve relationship between working-memory taxation and the effectiveness of DA tasks on memory, as proposed by Gunter and Bodner (2008). This theory forms the basis of the investigation by Littel and van Schie (2019), who investigated whether a linear dose relationship or an inverted U-curve dose relationship described the effects. They found evidence that there is a linear relationship between the amount of working memory taxation and memory degradation. They found that there may be an optimal level of working memory taxation. 

Both the Homer and Deeprose (2018) and Bartels et al. (2019) studies incorporated small and medium sample sizes (n = 38 and n= 80), respectively, with both utilising self-reported intrusive images. The studies were both pilots that sought to extend the EM/working memory interface beyond PTSD. Homer and Deeprose (2018) concluded that EMs reduce the vividness of voluntary images in social anxiety, but intrusive, involuntary images appear more resilient. Bartels et al. (2019) concluded that EMs reduce the self-reported emotionality, vividness and arousability of experience-based and imagination-based sexual fantasies across all participants. This offers an exciting direction for further research in the future to see if these findings are replicated in a clinical population.  

Limitations

There were limitations present in all the studies, the most significant of which was the choice of participants and the lack of a clinical population with a conventional diagnosis. In all six studies (Homer & Deeprose, 2018; Bartels et al., 2018; Mertens et al., 2019; Littel & van Schie, 2019; Eğeci & Özgün, 2019 and van Schie et al. (2019), the participants were identified as healthy, university students who received either course credits or a financial contribution for their participation. This leads one to question the implications for scalability regarding distinct clinical populations. Although laboratory studies are essential to explore concepts and dismantle critical components and elements, testing aversive memories within a clinical context is essential. 

The contribution of EMs to EMDR therapy has been the subject of debate for many years and has generated two systematic reviews and a meta-analysis that attempt to determine whether they form a crucial part of the therapy. Jeffries and Davies (2013) concluded in their systematic review that there was reasonable theoretical support for their inclusion in treatment, and Lee and Cuijpers (2013) found a significant, medium-effect advantage of EMs versus no EMs. Jeffries and Davies (2013) recommended that large-scale dismantling studies examining the difference in outcome between EMDR with EMs and EMDR without, as well as EMDR with other DA tasks, were needed.  Sack et al. (2016) undertook a randomised clinical trial comparing DA, EMs, and exposure only, with 139 patients diagnosed with PTSD. They used full protocol EMDR therapy with formally trained and clinically experienced EMDR therapists with adherence to the EMDR manual checked throughout the trial. Sack and colleagues concluded that there was no difference between the three conditions for treatment effects. All three treatment conditions led to a comparable, significant treatment effect and concluded that “The use of eye movements as a dual-attention task had no additional treatment effects compared to dual attention with visual fixation on a non-moving hand” (Sack et al., 2016, p.363).  

Regarding the robustness of the research, on a hierarchy of evidence, they represent a low level of evidence-for-practice as they are either descriptive or conceptual studies with limited sample sizes or populations (Day et al., 2017). Although these studies unquestionably contribute to the EMDR scholarly literature, they highlight the necessity for further robust research and clinical investigation. 

Conclusion

This paper has reviewed the current literature regarding the contribution of EMs to EMDR therapy and found that there is evidence that EMs reduce the vividness and/or emotionality of aversive memories in a non-clinical population (Homer & Deeprose, 2018; Bartels et al., 2018; Mertens et al., 2019; Littel & van Schie, 2019; Eğeci & Özgün, 2019). Laboratory studies have consistently demonstrated that when participants engage in two simultaneous tasks that use the same working memory resources, performance is reduced.  EMs impair the ability to hold a visual image in conscious awareness, resulting in the degradation of its vividness (Landin-Romero et al., 2018). Eğeci and Özgün (2019) found that EMs are an important and necessary component of EMDR treatment. Their study was the only one in this review that utilised a full EMDR protocol, whereas other studies have used DA tasks as a single element. In their study, Van Schie et al.  (2019) found that when the interventions (EMs) were longer, typically the same amount of time as a set of saccadic EMs in EMDR therapy, the result was a more significant reduction in the number of intrusions than in the control group.  

Jeffries and Davies (2013) reported that studies that cited using an EMDR treatment manual had a more significant effect size than those that did not, and Lee and Cuijpers (2013) found that treatment fidelity appears to increase the effects of the EMs/treatment. These findings suggest that the key to making meaningful changes within a therapy context is to stay faithful to the 8-phase protocol with highly trained and effective therapists, rather than isolating any single part of the overall therapeutic experience. The whole is greater than the sum of its parts.  

If EMs tax working memory and reduce vividness and/or emotionality, do they allow other aspects of the EMDR therapy to occur? Could EMs, therefore, be used in other trauma therapies, i.e., tfCBT, to aid recovery? Some of the data indicated that the additional effects of the eye movements might depend on the quality of the treatment delivery and that staying faithful to a protocol with trained professionals increases efficacy (Eğeci & Özgün, 2019). Therefore, further clinical research of EMs in a controlled clinical study using full protocol needs to be undertaken to understand the complex mechanism of action and to clarify the various processes at work. 

Clare Carter is a Registered Nurse (Mental Health) and Accredited EMDR Child & Adolescent Consultant and is currently a Clinical Nurse Specialist in a multiagency Youth Justice Service in Wales.